复杂装备退化状态早期识别方法研究

针对复杂装备早期退化状态难以识别的问题,提出一种将相关向量机(RVM)和Dezert-Smarandache 理论(DSmT)相结合的多特征融合决策识别方法。该方法首先分别采用时域分析法和时频域小波包变换法对装备的状态特征进行提取；之后将状态特征向量输入RVM模型中完成对状态属性的判定并获得各种状态模式的基本置信度分配；最后依据DSmT的PCR6规则对含有冲突信息的多个识别结果进行决策融合,得到早期退化状态的最终识别结果。在对某航空机电设备的实例应用中表明,该方法可以有效地解决信息高冲突条件下的早期退化状态识别问题,结果可靠准确。

Research on State Recognition Algorithm of Complicated Equipment with Early Performance Degradation

Aiming to the situation in which state recognition of complicated equipment with early performance degradation is hard to realize, a new fusion decision-making method based on multiple features extraction is presented, which compounds with RVM and DSmT. Firstly,a method, based on time domain analysis and wavelet packet decomposition, is used to extract the signal's feature separately; secondly, basic belief assignment function is constructed based on the output of the RVM model; lastly, PCR6 combination rule of DSmT is used to combine the different conflicting evidences and make the final decision. The application in a certain aerial electromechanical device suggests the approach is available to solve the problem of high-conflict information fusion when early vibration fault happens,and the recognition results are effective and reliable.